Connected data improves refinery

operations
Connecting data with a common knowledge management platform improves
visibility, response, and analysis to benefit refinery operations

JEFFREY ZURLO
GE Water & Process Technologies

T
he transition from board- create an ‘Industrial Internet of need for running miles of wires
mounted, pneumatic Things’. Connecting these data and cables. The increase in
refinery controls to elec- sources allows more rapid, sensors is already happening
tronic distributed control more transparent, and remote today – the average refinery
system (DCS)-centred control evaluation of system perfor- has seen a five-fold increase in
and data historian systems that mance, automated application the number of sensors
started to become popular in of proprietary analytic algo- deployed over just the last five
the 1970s has helped refinery rithms, and other system years.1 The trend to increase
operations improve produc- analyses that were impossible measurements via the use of
tion, reliability, and product to do previously. The increase sensors and other on-line
yield optimisation. However, in analysis and performance equipment is expected to
still today much of the data measurements provides signifi- continue for some time as the
originating from vendor service cant opportunities to better digitisation of industrial plants
companies remains locked optimise reliability and profita- provides another leap in
away in proprietary systems bility for these once isolated productivity and efficiency.
that are shared typically only systems. Robust data systems and
as periodic snapshots of perfor- There are several factors that large data storage capacity are
mance – collectively referred to are leading to this leap in tech- needed to handle the addi-
as ‘dark data’. Dark data in a nological capability. First, the tional data streams generated
refinery often includes infor- cost to produce sensors contin- as sensors and measurements
mation about the performance ues to drop and the variety of increase. Luckily, the cost of
of specialty chemical programs, operational and physical data storage has also dropped
equipment vibration monitor- parameters that can be meas- considerably to economically
ing, corrosion measurements, ured with sensors continues to accommodate the increased
inspection readings, and many increase. This allows more data load. Additionally, cloud
other ‘non-core’ areas of refin- sensors to be deployed in more based storage solutions are
ery operations. In the last places in a refinery, making becoming more popular daily,
several years, new sensors, visible what was previously it seems, opening up even
controllers, and data handling unmeasurable. In addition to greater storage and processing
infrastructures have been sensor technology, wireless capacity. While cyber security
under development to connect transmission has become is still a concern for refiners
this once dark data with the cheaper and more reliable, and industry,2 as security
traditional refinery operating allowing reduced cost of measures are improved there is
data as part of the movement installing sensors and main- a trend in the industry to allow
by dozens of companies to taining connection without the data to be transmitted ‘beyond

www.digitalrefining.com/article/1001309 PTQ Q3 2016 1

 The bottom line: connecting
Aviation 61 29 10 % dark data enables greater
insight to the system condi-
Wind 45 45 6 3%
tions, performance, and pitfalls
Power generation 31 63 6 % to better turn data to informa-
Power distribution 28 56 16 % tion and information to action.
Oil and gas 31 56 9 3%
Common connected data types
Rail 40 47 10 3%
• Real-time (streaming) data
Manufacturing 42 45 9 3% connection: making a constant,
Mining 24 55 18 3% always-on, connection between
the local sensors/controllers
Top/highest priority Within the top five priorities
and a cloud based control
Within the top three priorities Not a priority
system. This would provide
the most dynamic data stream
Figure 1 Industry survey on Big Data Analytics for analysis and action and
theoretically enable remote
the fence’ and stored remotely, investing in data scientists, closed-loop control of a system
thus making data available to programmers, and data analysts outside the refinery fence.
people and systems not physi- to provide the value that big However, this method is the
cally located at the refinery. data analysis can offer. More most difficult to keep secure
This movement contributes to sophisticated analysis due to from a cyber security perspec-
the tight integration of the better data visibility and tive and requires an absolutely
physical and digital worlds, advances in analytics helps to stable signal connection to
which many are calling the more easily anticipate impacts avoid dropped signals compro-
Industrial Internet of Things, of operational or feedstock mising system control.
and places us on the verge of a changes on system performance Additionally, utilising some of
new productivity revolution to and reliability. This field is the other methods described
transform industry on the scale growing rapidly in many areas below minimises the need to
of the industrial and digital of industry, so much so that have an ‘always on’ data
revolutions of the past.3 ‘data scientist’ was listed as the connection for the scope
According to a 2014 study by number one job in America for discussed in this article.
Accenture and GE, ‘Big Data 2016 by Glassdoor.5 It is quite • Near-real-time data communi-
Analytics’ is within the top possible that in the near future cation blends local closed-loop
three corporate priorities for data scientists will join the control and monitoring with
87% of oil and gas companies regular ranks of refinery profes- cloud based analytics and
surveyed and ‘not a priority’ sionals like engineers, performance monitoring. The
for only 3% (see Figure 14). instrument and controls teams, approach uses local, smart
The growing abundance of and reliability staff. In a recent devices to record sensor data
data provides great opportuni- interview with SAP, Peter and perform closed-loop
ties, but also additional Reynolds from the ARC control on systems. The system
challenges, in efficiently sifting Advisory Group argues that can periodically establish a
through large amounts of data adopting the technologies connection to the cloud based
and extracting valuable infor- provided by the Industrial platform and bilaterally trans-
mation to improve refinery Internet of Things could be a mit data between the local
performance. The emergence of key to survival for many oil and device and the platform. This
more advanced analytics is a gas companies by taking advan- increases data security since
growing trend in the industrial tage of optimising infrastructure the information is only trans-
world, including refineries, and to lower operating costs. “Why mitted in ‘bursts’, and
may indicate a shift in hiring monitor 50 pumps when you maintains closed-loop control
practices. Forward-looking can monitor 50 000 pumps locally, which is more stable.
companies have already started [from a centralised location].”6 This method still enables the

2 PTQ Q3 2016 www.digitalrefining.com/article/1001309

platform to monitor system Forum presentation, Sam note that this is a rapid-growth
performance rapidly and is Lordo outlined the need to area, so it is expected that the
usually sufficient to provide define key performance indica- scope of what can be achieved
feedback to out-of-control tors (KPI), key control using connected data will grow
conditions to enable timely parameters (KCP) and key rapidly over the next 5-10 years.
corrective actions. stress indicators (KSI) as essen- We are only very close to the
• Edge computing: a variant on tial elements of a well-defined starting line today.
the above near-real-time performance management
communication and a growing system.7 For a system to be Improved oversight and
trend, it increases the intelli- valuable to refinery operations, communication - automating
gence of the local smart device a performance management performance assessment and
to enable more complex analytic and communication system dashboards
execution locally and autono- needs to be able to not only With a connected data system
mously. This approach takes identify when a system is out everyone looks at the same data
advantage of the greater of compliance but also provide pool. Plant personnel, vendor
processing power and lower a simple mechanism to ensure field teams, and corporate
power requirements of today’s action is taken to bring a subject matter expert staff in
computer processors to unload system back into control. A both the refinery and vendor
communications and server common platform utilising organisations have visibility to
bandwidth while enabling connected data can provide a system status and data at the
controls beyond traditional single data view that is trans- same time, from the same
PID-type control schemes. Data parent, can automate system. This breaks the historic
transparency and system specialised analytic analysis to linear communications chain
performance are preserved by allow more advanced perfor- from the past that slows
communicating critical data and mance assessment, and can communication and often
analytic outputs to the platform, allow for defined control limits results in some parties using
while pushing more ‘data for KPIs, KCPs, and KSIs. This out-of-date information (see
crunching’ down to the local type of comprehensive system Figure 2). Also, dashboard
level. provides a clear mechanism to ‘health checks’ are requested
• Manual, or on-demand, data turn data to information, and more frequently by plant and
upload: intended mostly for information to corrective corporate technical personnel to
local data obtained on a spot actions. provide visibility to current
basis, such as laboratory results Utilising a system that allows system performance and antici-
for field readings not captured all the pertinent data to be pate any potential future
by on-line sensors or instru- collected on one platform and operating issues, often needing
ments. Data of this type is available to a wider group of daily updates. A connected data
updated either on some set support personnel has several system can accommodate crea-
frequency, or as it is generated advantages for the refinery. tion and maintenance of
by a person. There are several Pairing connected data with the dashboard systems for any level
mechanisms to incorporate this increased scope of on-line of organisation. In addition,
data, such as manual template sensors and increasing capabili- analytic outputs and data feeds
upload, live forms, or file ties of local smart devices from vendor measurement and
transfer/email. Historically, benefits refinery operations and control systems can be directed
this has been some of the most reliability at a number of differ- back into the plant’s DCS or
difficult data to incorporate ent levels. The following data historian system. This
into the data stream for refin- sections will describe the main enables expansion of the data
ery operation and is therefore a types of advantages seen so far typically captured in the plant’s
huge, mostly untapped in refinery operations and historian to more detailed
resource of information. include examples that highlight vendor information, preserves
how these improvements were the ability for plant personnel to
Advantages of connected data achieved and the value be able to see all their operating
In a 2012 AFPM Technology provided to the refiner. Please data in one application, and still

www.digitalrefining.com/article/1001309 PTQ Q3 2016 3

 take advantage of the improved
Traditional communications that connected
data systems provide.
Field devices Refinery data Local readings
Tighter system control through
faster recognition and response
to out-of-control conditions
Triggering alarms from the
Performance Account team Analytics automated data gathering and
performance assessment of
KCPs and KPIs against estab-
lished operating limits means
that loss of control is recog-
Management / Regional nised as it happens, not at the
Operations
Engineering engineering next operator or vendor visit to
the system. While this is noth-
ing new for a refinery’s core
operating and control functions
Corporate Corporate
SME SME that feed through the DCS and
Technology
data historian systems, vendor
systems often use proprietary,
disconnected systems and spot
Connected data data joined with plant opera-
tional data to assess system
Smart devices Refinery data Local readings performance. Historically this
has been done via pooling of
this data locally using spread-
sheets and local analytic
packages, which is not very
Cloud based
platform efficient and sometimes intro-
duces significant lag time
between data generation, issue
Analytics recognition, and corrective
Account team Remote actions.
centre
Performance
Case 1: reducing upset impact
through rapid response protocols
This example highlights the
Management / Regional
Engineering
Operations
engineering
advantage of connected data in
greatly improving response to
an event-driven, upset condi-
Corporate Corporate tion. A refinery cooling tower
SME SME
Technology utilises local automated closed-
loop control of the cooling
water chemistry and on-stream
Refinery data monitoring of several operating
parameters such as pH, chlo-
Common platform rides, conductivity, and
chemical residual testing. Due
to a valve misalignment
outside the battery limits of
Figure 2 Changing communication avenues this operating unit, concen-

4 PTQ Q3 2016 www.digitalrefining.com/article/1001309

trated caustic was fed to the
14.0
make-up line of this cooling
tower instead of the normal 12.6
treated surface water.

CT 5 supply pH
(bleach skid)
Replacing water in the 11.2
make-up line with caustic over-
whelmed the control system, 9.8
resulting in an almost immedi- HH
8.4 H
ate increase in pH of the tower L
LL
above 12 pH (see Figure 3). 7.0
Sensing the system was out of

12

12

12

12

12

12

12

12

12
0

0
control, the cloud based plat-

/2

/2

/2

/2

/2

/2

/2

/2

/2
/9

/9

/9

/9

/9

10

10

10

10
22

24

26

28

30

2/

4/

6/

8/
form issued an alarm to the
local account team via email, Figure 3 Cooling tower pH measurement
text, and direct system alert.
The account team was on-site brought the system back the problem. The function of a
and immediately contacted the under control in a shorter time remote monitoring centre is
unit operators, who were than would have otherwise similar to the operations central
unaware of the issue, as a field been possible, and good plant control room staff at a refinery,
inspection round had been communication enabled quick but focused on the systems the
completed just before the inci- identification and correction of vendor helps to service and
dent occurred. The local the root cause to remove the generally do not have direct
vendor team and the operators caustic from the make-up control on the process manipu-
responded to the condition by water line. A high pH event of lated variables, just visibility to
verifying the condition was this magnitude in a cooling the system conditions. Also,
real via hand measurement at system would typically result support staff can more easily
the cooling tower basin, in widespread scaling of heat and more routinely access the
making aggressive system exchange equipment and information for a given
changes to reduce the pH of reduced unit run length. This refinery, typically through a
the cooling water back to the would cost a refinery signifi- secure password-protected web
normal control range to cant lost opportunity dollars portal. This allows additional
prevent scaling, and communi- from energy efficiency loss or oversight by people with the
cating to other plant personnel unplanned outages, but was right skill set to supplement
to identify and correct the root avoided by quick event recog- the expertise provided by the
cause of the problem. Since nition and rapid response to local service and support team.
cooling water pH was identi- get the system back under
fied as a critical alarm for this control. Case 2: identifying and correcting a
system, the vendor’s remote chronic operating issue via pattern
central monitoring centre was Increase the layers of recognition
also directly alerted by the protection through additional Shortly after commissioning a
system to ensure that the local monitoring and analysis by off- continuous, on-line chloride
personnel and customer were site personnel analyser in a crude unit atmos-
aware and corrective actions Connecting data on a cloud pheric tower overhead system
were being taken. based platform allows the crea- at a North American refinery,
As a result of the systems in tion of a remote monitoring one of the vendor’s off-site
place, the upset condition was centre or centres, staffed by regional engineers noticed a
identified in minutes instead vendor personnel to provide an repeated pattern of high chlo-
of hours; it would have been additional layer of oversight on ride excursions on an otherwise
several hours until the opera- system performance and act as fairly well controlled system. As
tor was scheduled to a command centre during can be seen in Figure 4, these
physically monitor the cooling system upsets to help ensure excursions were occurring once
tower again. Rapid response local personnel are aware of or twice a week, have very

www.digitalrefining.com/article/1001309 PTQ Q3 2016 5

 tion rate using the overhead
155
chloride measurement and to
COMS unit overhead 124
reduce sodium load in the
chlorides, ppm reduced crude and vacuum
91 residuum from the crude unit
fed to the FCC unit.
62

31
HH Expanded analytics improve
H insight to system performance
0 L Proprietary analytics previ-
ously required someone to
14

14

14

14

4
01

01

01

01

01
20

20

20

20

/2

/2

/2

/2

/2
collate and organise data from
1/

1/

1/

1/

12

12

12

12

12
/1

/1

/1

/1

1/

2/

3/

4/

5/
27

28

29

30

a variety of sources, input data

Figure 4 Overhead chloride excursions to run the analytic, and then
analyse the output. A common
data platform can automate
155
this process whenever applica-
COMS unit overhead

124 ble new data is received into

chlorides, ppm

the connected platform, allow-

93 ing vendor and plant personnel
to focus on interpreting the
62
outputs and engineer correc-
HH
31 tive action plans instead of
H
spending time in data handling
0 L
09:18 09:20 09:22 09:24 09:26 09:28
activities. Additionally, a
09:19 09:21 09:23 09:25 09:27 09:29 common data handling plat-
Time, h:m form that uses appropriate
timestamp protocol eases the
Figure 5 Overhead chloride in control ability to export synchronised
data when the rhythm of data
sharp peaks and a quick recov- achieve targeted dilute caustic acquisition is not the same
ery. This system had a history strength. Adding the water between data parameters, for
of salt-induced corrosion and first caused injection of essen- instance pairing continuous
fouling, resulting in extensive tially pure water into the crude operating flows with periodic
equipment damage and short preheat and resulted in a loss laboratory testing. This
equipment life, unplanned shut- of chloride control in the over- improves usability of the data
downs, and high maintenance head until the batch blending for analysis and review.
costs, so the short-term spikes was completed. Although this
were very concerning and a practice had been in place for Case 3: improving waste treatment
root cause investigation was many years, it was not identi- plant performance
launched. fied as an issue until Refinery wastewater treatment
The root cause was identi- continuous chloride measure- units often are the areas with
fied: the operating procedure ment and visibility to the data the darkest data. Many refiner-
used for many years to blend a stream was achieved. As a ies that have moved to DCS
new batch of caustic injected to result of the investigation, the systems decades ago still use
control overhead chloride plant now purchases pre-di- manual log sheet entries to
concentration. In brief, when luted caustic to avoid periodic capture wastewater treatment
the tank inventory was low, a spikes (see Figure 5). As a side data. This limits the extent of
set amount of make-down note, the refinery has further data analysis and can impede
water was charged to the caus- increased control of caustic troubleshooting efforts during
tic batch tank, followed by a injection by implementing root cause and performance
concentrated caustic feed to closed-loop control of the injec- optimisation activities.

6 PTQ Q3 2016 www.digitalrefining.com/article/1001309

 An exception is a refinery on
the US Gulf Coast that had 600 90

Organic acid to desalters, gpd

Organic acid to desalters
transitioned to using its data
500 Clarifier bed depth 80

Clarifier bed depth, in

historian system to capture and
store the operational data at its 400 70
waste treatment unit. Thus,
when the refinery started to 300 60
share the data with the treat-
200 50
ment vendor’s cloud based
knowledge management 100 40
system, it was easy to include
the data captured at the waste 0 30
treatment plant as well.

14

4
14

14

14

4
14

14

14

15

15

5
01

01

01

01

01
20

20

20

20

20

20

20

20

20
/2

/2

/2

/2

/2
This proved to be very useful

0/

1/

1/

1/

2/

2/

2/

1/

1/
11

12

/1

/1

/1
/1

/1

/1

/1

/1

/1

/1

2/

9/
16

23

30
7/

5/
31

14

21

28

12

19

26
when the waste treatment area
started experiencing significant
but intermittent operational Figure 6 Desalter to waste treatment correlation
issues in the activated sludge
clarifier area. The biological clarifiers, it was difficult to developed at the waste treat-
system was generating an detect the correlation using ment plant to better handle
excessive amount of bulking just historical trend analysis. high organic acid feed rates
bacteria, swelling clarifier beds Organic acid programmes are that involved more aggressive
and preventing effective sludge used at the desalter to improve operating changes and use of
settling. This caused difficulty desalter performance in the acid injection dashboard
meeting waste treatment final response to a need for data as an ‘early warning’
effluent water quality, which controlling desalter pH due to system. As a result of
required emergency measures increased desalter pH and improved operational control
to be taken to avoid environ- contaminant levels at the crude and increased communication
mental penalties. units as a result of expanding across operating areas of the
The vendor’s corporate refin- the crude diet to process tight refinery, waste treatment clari-
ery wastewater SME, in oils and other crude oils not fier operation has improved
addition to the local vendor historically processed at this significantly and is no longer
personnel and refinery staff, refinery. Due to the strong considered an operational
investigated the issue to deter- economic incentive to process bottleneck, preserving flexibil-
mine root cause and corrective these opportunity crudes, a ity in operation and crude oil
actions. Using the connected solution that would preserve purchases and maintaining
data system allowed rapid crude oil flexibility and good quality brine from the
application of multivariate improve waste treatment oper- desalters.
regression analytics for data ations was desired.
associated with refinery opera- As a result of the analysis, Improve integrity and
tions and the waste treatment tighter control of acid feed was economic operating windows
performance. Among all the initiated at the crude units to Crude tower operations are
operating data considered, a limit total stress on the waste faced with several constraints,
strong correlation was identi- treatment plant and communi- including operating windows,
fied between injection of cations were improved by economic demands, and
organic acids at the crude unit adding total organic acid feed several factors that can affect
desalters for pH and contami- rates to the daily dashboard reliability and efficiency. Using
nant control with the increase reporting system. When high connected data and taking
in clarifier bulking (see Figure injection rates at the crude advantage of Big Data analysis
6). Due to the large lag time units are necessary based on techniques, an analyst can now
between the injection at the crude unit conditions, an oper- develop more dynamic analytic
desalter and the effect at the ational response plan was analyses to simultaneously

www.digitalrefining.com/article/1001309 PTQ Q3 2016 7

 expand operational flexibility included those injected into the Further, as the operating objec-
whilst maintaining mechanical refinery for neutralisation, tives and crude diet change
integrity within the bounds of introduced as contaminants in over time, it can now rapidly be
these constraints. The analyst the crude oil stream, and recy- determined when the system
should be able to more easily cled and degraded amines leaves safe operating conditions
identify the operating points to from various refinery circuits. and moves into a higher salting
maximise production goals, A number of actions were risk profile. When this occurs,
such as producing more diesel implemented to reduce or elim- the relative ranking of all the
fuel, while avoiding conditions inate the problematic amines drivers responsible for the
that impact reliability, such as and chlorides contributing to excursion can be easily and
crude tower overhead corro- salting to reduce salting from a clearly identified, providing a
sion and salting. Further chemistry perspective: the path to most effectively mitigate
system improvements can also desalter operation was modi- the risk and re-establish safe
be made by identifying correla- fied to improve crude oil operating conditions.
tions and root causes, and contaminant removal, and the Fractional salting requires
opens up the ability to perform refinery steam amine several operating parameter
predictive analytics on refinery programme was changed to inputs in addition to what is
operations. Simply put, utilis- reduce impact on overhead salt normally considered for tradi-
ing connected data and formation. tional salt point analysis, such
advanced analytics can help to In parallel, an analytic routine as additional crude column
better explain the state of a was implemented to better conditions, product fraction
system, determine when action quantify the time salt formation flow rates, pumparound rates
is needed to improve the is occurring in the system, and physical conditions like
system, and define the factors resulting in an output variable temperatures and pressures.
that are the main drivers to coined as ‘Fractional Salting’. Since this refinery had already
affect positive change. Whereas traditional techniques been providing data to the
only allowed salt point compu- vendor’s cloud based system,
Case 4: improving reliability by tation at specific points in time, expanding the data sent to run
improving integrity operating the new analytic toolset enables the analytic was fairly straight-
window definition visualisation of the entire ‘salt forward. This provides several
A North American refinery potential’ for the system, its potential benefits to the refiner,
was experiencing reliability sensitivities to all input varia- including being able to:
issues in a crude unit overhead bles, and their changes over 1. Improve visibility to the risk
system. Fouling and corrosion time. Using the fractional salt- of salting corrosion and fouling
impacted unit reliability, caus- ing approach, one is able to by moving from spot analysis
ing short equipment life and dynamically calculate the frac- to describing the boundaries of
increased unplanned mainte- tion of time a system is in a the salt formation given the
nance costs. A solution was salting condition based on both variability in operating param-
desired that identified the root a given set of operating condi- eters, not just current data or
causes and helped to determine tions and the variability of each averages. Figure 7 shows the
effective corrective actions. operating parameter. This calculated salt points for a
Traditional techniques provides a method to rapidly particular operating condition
deployed identified salt-in- assess the current state of the at this refinery. The highlighted
duced fouling, under deposit overhead system in terms of point is the operating point at
corrosion, and high overhead salting, quantitatively deter- the time of measurement and
velocities as root causes. Amine mine if further improvements the ‘cloud’ of data points
speciation testing revealed over are needed, and numerically around it indicates the amount
a dozen types of amines pres- show the level of improvement of variability seen during an
ent in the overhead system that can be achieved by changes operating mode. In this case,
originating from multiple made on the system to reach the fractional salting was calcu-
sources contributing to salting. specific goals within the lated at just over 40%, with
Sources of the problem amines constraints of the system. most of the operating points

8 PTQ Q3 2016 www.digitalrefining.com/article/1001309

being either in the salting zone
150
(orange area) or below the
margin of safety for salting
(beige band). This is an area of 140

concern for salting and repre-

sents a high potential for 130

OHD_T
negative impact on reliability.
In fact, the plant confirmed 120
that overhead fouling and
corrosion did take place during 110
this period of operation. Note
that the chart compares over- 100
head temperature to reflux 1000 2000 3000 4000 5000 6000 7000
Reflux
rate, but the analytic calculates
salt formation potential scalar
fields around the variation of Figure 7 Base case: salting condition, 40% fractional salting
all considered operating
parameters.
150
2. Calculate the sensitivity that
changing any specific opera-
tional handles, both average 140

value and variation, makes on

fractional salting. This can be 130
OHD_T

used to advise on changes to

improve reliability, improve 120
yield profile, or both. Figure 8
shows the salting potential after 110
changing the operating point
and improving variability of 100
several operating parameters in 1000 2000 3000 4000 5000 6000 7000
Reflux
the system. As can be seen
graphically, the potential for
salt precipitation has been Figure 8 Improved case: fractional salting reduced to 6.5%
significantly reduced with the
operating point moving into the produced. This allows advisory would otherwise go unnoticed,
safe zone (green area) and the control for the crude unit over- drives visibility to the current
spread of data points shrinking. head, to enable meeting state of system performance,
Specifically, the concentration of operational goals while avoid- and clearly identifies efficient
diglycolamine (DGA) in the ing negative impact to paths toward optimised
overhead was reduced and reliability due to salt systems to improve yield
tighter control of several operat- formation. profiles and preserve system
ing parameters to reduce integrity. As this field grows in
variability improved fractional Conclusion industry, and specifically in
salting from over 40% to 6.5%. Connecting data with a refineries, new opportunities
3. Given a set of defined common knowledge manage- for data scientists and analytics
constraints around any or ment platform allows are emerging, with the poten-
several operating parameters, improved visibility, accelerates tial to change the face of
calculate an optimised case that response, and enables deeper refinery operations and usher
can achieve multiple goals, analysis to improve refinery in a new era of achieving oper-
such as reducing fractional operations, reliability, and ational calm in the face of
salting below 5% while maxim- profit potential. Bringing dark increasing variability and
ising the amount of distillate data to light reveals trends that tighter regulations.

www.digitalrefining.com/article/1001309 PTQ Q3 2016 9

 Acknowledgements industrial-iot-an-opportunity-for-fast- Sam Lordo, Nalco Champion, 2012 AFPM
The author would like to recognise the moving-innovators-a-threat-to-slow- Q&A and Technology Forum.
significant contributions of Dr Collin moving-dinosaurs Jeffrey A Zurlo is Senior Strategic
Cross, David Mason, Henry Grichuk, and 4 Industrial Internet Insights Report for Project Manager with GE Water &
Steve Bakas for their help in providing 2015, Accenture, 2014, www.accenture. Process Technologies. With over 25
data and feedback for this work. com/us-en/_acnmedia/Accenture/ years’ experience in process engineering,
next-gen/reassembling-industry/pdf/ refinery treatment sales and services,
References Accenture-Industrial-Internet-Changing- hydrocarbon product applications and
1 Gonzalez R, Underutilization of refinery Competitive-Landscape-Industries.pdf technical support, he provides marketing
automation systems, Fuel Magazine, Dec 5 The 25 best jobs in America right and technical support to the refining
2015, 39. now, according to employees, Jacquelyn and gas processing industries. He holds a
2 Big Success with Big Data, Accenture, Smith, Business Insider, 20 Jan 2016, BE in chemical engineering from Stevens
April 2014, www.accenture.com/us- www.businessinsider.com/best-jobs-in- Institute of Technology, New Jersey.
en/_acnmedia/Accenture/Conversion- america-right-now-2016-1
Assets/DotCom/Documents/Global/ 6 Low Oil Prices Prime Oil and Gas
PDF/Industries_14/Accenture-Big-Data- Industry for IoT Adoption, John Ward,
POV.pdf Forbes Business, 1 Feb 2016, /www.
LINKS
3 Industrial IoT: An Opportunity for forbes.com/sites/sap/2016/02/01/low-
Fast-Moving Innovators; a Threat to oil-prices-prime-oil-and-gas-industry- More articles from the following
Slow-Moving Dinosaurs, Industrial IOT/ for-iot-adoption/#469b81a75ce2 category:
Industrie 4.0 Viewpoints, 6 bJan 2016, 7 Key Performance Management – A Instrumentation, Automation and
http://industrial-iot.com/2016/01/ Comprehensive Look At Performance, Process Control

10 PTQ Q3 2016 www.digitalrefining.com/article/1001309